Governmental power sharing: imperative?: A comparative case study on decentralization and its influence on public support for secessionism

Governmental power sharing: imperative?

A comparative case study on decentralization and its influence on public support for secessionism

Andrés Durante

Master’s Thesis

Uppsala University, 2020

Department of Government

Supervisor: Moa Mårtensson

Abstract

Scholarly debates about the merits of decentralization have long been central in comparative politics. However, there is a lack of consensus in the existing literature on its relationship with secessionism, and previous scholarship on autonomy and power sharing suffers from several shortcomings that make it difficult to evaluate the influence of decentralization on secessionist conflict. To address this research gap, a new theoretical framework is proposed which suggests inclusion for regional elites within the state’s national decision-making institutions is imperative for the prevention against secessionism in established democracies. Utilizing an innovative comparative process tracing method, a theory-driven within-case empirical analysis and between-case comparison is conducted between Spain and Catalonia, and the United Kingdom and Scotland. The findings illustrate that decentralization with a higher degree of governmental power sharing (shared rule) is associated with a lower degree of public support for secessionism, and this association is produced through a social mechanism of power.

Keywords: decentralization, power sharing, secessionism, Spain, Catalonia, UK, Scotland

Table of contents

LIST OF FIGURES AND TABLES ...

1 INTRODUCTION ... 1

2 THEORY ... 3

2.1 Literature review... 4

2.2 Theoretical framework ... 7

3 METHOD ... 10

3.1 Operationalization of dependent and independent variables ... 10

3.1.1 Dependent variable: public support for secessionism ... 10

3.1.2 Independent variable: decentralization... 11

3.2 Case selection ... 13

3.3 Control variables ... 17

3.4 Comparative Process Tracing (CPT): a two-step methodological approach ... 19

3.4.1 Within-case step... 20

3.4.2 Between-case step ... 23

3.5 Empirical material and data ... 25

3.6 Validity and reliability ... 25

4 WITHIN-CASE EMPIRICAL ANALYSIS... 26

4.1 Spain and Catalonia ... 26

4.1.1 Overview of decentralization and public support for secessionism ... 26

4.1.2 Implementation phase (a); 1978-1980 ... 28

4.1.3 Interpretation ... 29

4.1.4 Bargaining phase (b); 1980-2010 ... 30

4.1.5 Interpretation ... 33

4.1.6 Mobilization phase (c); 2010-present ... 34

4.1.7 Interpretation ... 37

4.2 United Kingdom and Scotland ... 37

4.2.1 Overview of decentralization and public support for secessionism ... 37

4.2.2 Implementation phase (a); 1997-1999 ... 39

4.2.3 Interpretation ... 41

4.2.4 Bargaining phase (b); 1999-2011 ... 42

4.2.5 Interpretation ... 45

4.2.6 Mobilization phase (c); 2011-present ... 45

4.2.7 Interpretation ... 47

5 BETWEEN-CASE COMPARISON ... 48

5.1 Implementation phase (a) ... 48

5.2 Bargaining phase (b) ... 49

5.3 Mobilization phase (c)... 49

6 RESULTS ... 51

6.1 Consistency with the hypothesis ... 51

6.2 Limitations ... 51

6.3 Alternative explanations ... 51

7 SUMMARY AND CONCLUSION ... 53

REFERENCES ... 56

APPENDIX ... 67

List of figures and tables

FIGURE 1. Theoretical argument... 8

FIGURE 2. Theoretical framework... 9

FIGURE 3. Reverse causation ... 52

TABLE 1. Controlled comparison... 14

TABLE 2. Theoretical social mechanisms in a decentralization process ... 22

TABLE 3. Public opinion on Catalonia’s political status within Spain ... 27

TABLE 4. Public opinion on how Scotland should be governed... 39

"Seré curioso / señor ministro / de qué se ríe / de qué se ríe / (...) Aquí en la calle / suceden cosas / que ni siquiera / pueden decirse"

Mario Benedetti

1 Introduction

Public support for secessionism is on the rise throughout the world. Eight movements pursued an independent state in 1915, in comparison with 59 in 2015 (Fazal 2018). To make matters worse, consolidated democracies are not immune to secessionist tendencies. The Scottish independence referendum held in 2014, the current implications of Brexit on Scotland and the UK’s national cohesion, the Catalan independence referenda held in 2014 and 2017, and the eruptions of violence experienced in Spain after the sentencing of nine Catalan leaders in 2019 are all recent examples of manifestations provoked by underlying secessionist tensions in modern decentralized states with regional minority populations. Scholarly debates about the pros and cons of decentralization have long been central in comparative politics. However, despite the ongoing secessionist growth, there is little consensus in the existing literature on the relationship between autonomy concessions and secessionism (see Bakke 2015). Some argue that decentralization can contain conflicts of self-deterministic character by meeting regional minorities halfway through institutional channels, allowing the voicing of their demands.

Others suggest that decentralization fuels further conflict and promotes state disintegration.

Specific power sharing elements of decentralization originally formulated by Elazar (1987, 1991: 8) have been scrutinized in studies. His distinction of self-rule and shared rule – the capacity of regional governments to autonomously exercise authority over those who live in its territory, and the capacity to co-determine the exercise of authority for the country as a whole, respectively – has provided researchers with conceptualizations that are useful to help understand the intricacies of decentralized governance. However, analyses looking on regional autonomy in conjunction with power sharing have rarely been made, and few scholars have considered the effects of self-rule in combination with shared rule (Cederman et al. 2015: 356).

An important research gap is identified as several contrasting findings suggest that decentralization “can have a complex and multifold effect” (Massetti and Schakel 2017: 433).

The following research question was developed to seek further evidence on this phenomenon:

“How, and through what mechanisms, does decentralization with a higher degree of

governmental power sharing (shared rule) influence public support for secessionism?”

This question is of importance because it seeks to understand how decentralization reforms can be an effective instrument to manage and contain the challenges that secessionism poses to the territorial integrity of states. This is of particular concern for policy formulation and implementation, as autonomous claim by regional minorities are of great relevance for many states today. It also distinguishes itself from the existing debate about peace-preserving decentralization by relocating research direction. The intent is to explore how decentralization influences an outcome of secessionism in one direction or the other rather than asking whether decentralization ameliorates or exacerbates secessionism in the first place. More specifically, the aim is to explicitly study the mechanisms of a decentralization process that are either secession-preventing or secession-inducing in a given context. The revision of directional focus is made because ultimately, the existing debate is inconclusive. There is plenty of empirical evidence with credible argumentation to support any of the two sides. Also, the diverse accounts of conflict in states which are decentralized, where adverse outcomes have been generated by decentralization reforms, have not been adequately addressed.

The thesis is structured like this: a theory section is introduced, where theoretical concepts are

defined, and a review of the existing literature is made. Then, the study’s theoretical argument

is described, together with the theoretical framework and hypothesis. A methodological section

follows where operationalizations are outlined, the case selection strategy is illustrated, and

method of choice is presented. Cases are subsequently empirically analyzed and compared

between each other, followed by a discussion and presentation of results, limitations and

alternative explanations. To conclude the study, findings are briefly summarized, result

implications on theory are discussed, and recommendations for future research are suggested.

2 Theory

Numerous definitions of decentralization can be encountered in the literature. However, some consensus among scholars can be identified. Lijphart (1999: 175) recognizes the concept as a framework constituting of an elemental feature that guarantees a power distribution among central and regional levels of government. Accordingly, this study defines decentralization as a governmental system where tiers of government share power, and where lower tier units exercise some form of self-rule. This definition is only legitimate in democratic states. This is because as Brancati (2006: 652) explains, the accreditation of power to regional legislatures can principally be done by non-democracies. In practice, however, regional leaders with pro- government sentiments are regularly installed, so as to not challenge the authority of the central state. As such, regional governing powers are fundamentally disregarded or violated.

Federalism is commonly believed to be analogous to decentralization. Therefore, it is of importance to stress the theoretical differences between them. Lijphart (1999: 176) illustrates how the concepts relate with each other by highlighting the primary characteristics of federalism; non-centralization and decentralization of power. What distinguishes federalism are its secondary characteristics, which include a bicameral legislature with a strong federal chamber, a written constitution, and a supreme court or special constitutional court. King (1982:

74) does not consider the concept of federalism institutionally, but rather views it as an ideological philosophical principle of self- and shared rule. Burgess (2006: 2) argues that such a federalist principle is in support for a federation and serves to recommend and promote such an entity; that is, a federal state. Watts (1996: 8) explains that federalized states have a central government in combination with strong constituent units, where citizens directly elect its leadership by exercising power, and this power is authorized by a constitution.

A desire to consider de facto decentralized states which do not define themselves as federal, but share important characteristics generally attributed to federalism, has meant scholars are increasingly using the term decentralization instead of the term federalism (Brancati 2006: 654).

The concept of “power sharing” is an essential component in decentralized governance. As

such, it requires terminological attention. In this study, it is used to characterize any scheme of

governance that permits group representatives (elites) to engage in shared decision-making processes – either by granting territorially-concentrated groups regional autonomy (which is labeled as territorial power sharing or self-rule), or through inclusion within the national decision-making institutions (which is labeled as governmental power sharing or shared rule).

Governmental power sharing is closely related to the concept of consociationalism established by Lijphart (see Cederman et al. 2015). The definition used in this study primarily concerns decentralization within unitary states however and is therefore limited to a greater extent.

In this study, Hechter’s (1992: 267) definition of secessionism is used, namely “a demand for formal withdrawal from a central political authority by a member unit or units on the basis of a claim to independent sovereign status”. It is the fundamental desire of territorial group members and elites for an independent state. Secession is successful when the withdrawal is recognized by the host state and by others in the international community. It differentiates itself with separatism, which does not aim for neither formal withdrawal nor recognition (Ibid).

In the next section, existing literature on the relationship between the concepts of decentralization and secessionism is reviewed.

2.1 Literature review

At large, research on decentralization and secessionism has produced two contradictory

positions; a pessimistic and an optimistic. Scholars adherent to a more pessimistic point of view

argue that, as a method of conflict resolution, decentralization is ineffective and a hindrance to

peace (see Brubaker 1996; Kymlicka 1998). Chapman and Roeder (2007) contend that

decentralization can strengthen ethnic identities through the formal recognition and

legitimization of individual ethnic groups. Elkins and Sides (2007: 693) sustain that

decentralization provides considerable resources which can be used by groups to pressure the

state. Cornell (2002: 251-252) believes autonomous regions are inherently problematic, as they

capacitate ethnic groups to challenge central states’ authority through subnational political

institutions. By promoting ethnic mobilization and facilitating insurgency to maintain or

increase regional autonomy within territorial communities, decentralization undermines

interethnic stability. Self-governance ultimately leads to conflict, secessionism, and state

dissolution according to Cornell (2002: 275-276). Roeder (2009: 214) shares similar reasoning,

arguing that by preserving a sovereign central state which also responds to autonomous claims of ethnic groups, decentralization supports an inherent incompatibility. The fundamental issue of the belonging of ethnic minorities in a common state is not sufficiently addressed. Instead, dangerous implications for potential state disintegration are maintained by this vulnerable equilibrium. Roeder (2009: 203, 208) argues that decentralization increases the risk of secessionism by causing political vulnerability as actors become enclosed between two perils:

centralization and dissolution. Experimenting with institutional power sharing arrangements is unlikely to resolve the fundamental issue. Chapman and Roeder (2007) advocate for partition as a preferred solution over power sharing and autonomy.

Other authors hold that decentralization is related to the emergence and aggravation of mobilization (e.g. Bunce 1999; Hale 2000; Kymlicka 1998; Treisman 2007). The devolution of power and creation of regional self-governing institutions are catalysts for generating opportunities that regional elites use to demand further authority. Scholars contend secessionist movements mobilize public support to pressure central governments to agree to widen the jurisdiction of regional institutions, or to pursue outright independence.

Scholars advocating a more optimistic viewpoint argue that by successfully responding to autonomous demands via an increase of self-governance, decentralization decreases the probability of ethnic conflict and secessionism. Hechter (2000: 8-9) believes decentralization provides central governments with cost-effective measures to maintain state sovereignty by granting autonomy concessions for regional ethnic groups. Thus, it is an appropriate system of government for ethnically divided states. Lustick, Miodownik and Eidelson (2004: 210) contend regional autonomy reduces conflict by ensuring governments take into consideration the most vital concerns of minorities. Decentralization appeases potential secessionist groups by providing them with adequate representation and protection. Representation implies a legal framework with formal recognition and political integration (Bakke 2015). Other authors argue that decentralization diminishes the probability of mobilization (e.g. Gurr 2000; Lijphart 1994;

Stepan 1999; Tsebelis 1990). Political institutions which facilitate the expression of local grievances and have the ability to articulate immediate responses to collective needs decrease contentious motivations of regional actors by increasing loyalty towards the central state.

Evidently, the relationship between decentralization and secessionism is controversial. Hechter

and Okamoto (2001) summarize it as “remarkably murky”. Overlaps between the two stances

in the literature have been attempted but the results differ. For example, Mattes and Savun (2009) and Walter (2002, 2009) endorse power sharing for conflict resolution but remain skeptical about autonomy. In contrast, Jarstad and Nilsson (2008) argue that autonomy restrains conflict more effectively than governmental power sharing (see also Martin 2013).

Theories concerning ethnic conflict and its relationship with secessionism have traditionally focused on grievance-based arguments, although much of recent literature within the subject has cast doubts on said reasoning (see Blattman and Miguel 2010). They originate from research illustrating how ethnic rebellion is instigated by governments maltreating minorities (e.g. Gurr 1993, 2000; Petersen 2002). Conflicts of secessionist character are often caused by ethnic minorities competing for state power. Thus, the state plays an integral part, as ethnic groups lacking access to state power are more likely to pursue independence (Gurr 2000: 195; Erk and Anderson 2009: 192). Hechter (2004) presents support for an increased possibility of ethnonationalist conflict in multiethnic states when ethnic minorities do not experience high levels of self-governance. Similarly, Girardin et al. (2015) argue that “[group] exclusion from government is associated with a higher risk of civil war”.

According to Bakke (2015: 13-15), self-determination conflicts revolve around demands of independence, decentralization reforms, or greater autonomy from central governments (central elites), regional minority groups (regional elites), and the population they represent. Such demands are either ameliorated or exacerbated by political institutions which govern the relationship between central and regional elites. She believes state institutions run the risk of institutionalizing regional and ethnic identities, supplying assets for mobilization, and encouraging the forming of regional and ethnic based political parties which have been shown to increase the risk of ethnic conflict and secessionism (see Horowitz 1985; Brancati 2006).

Bakke (2015: 16-17) argues that autonomous institutions which are isolated from the state’s central decision-making institutions are more susceptible for mobilization by regional elites.

She asserts that strong political ties between central and regional elites decreases the risk of secessionist conflict substantially. Cederman et al. (2015: 358-359) claim that excluding ethnic groups from political influence in the state’s central executive is associated with a higher risk of ethnic conflict and secessionism. The risk is amplified considerably if the central government implement unfavorable policies to ethnic minorities or attempt to recentralize power after

1

Important to note that these findings involve effects of types of power sharing adopted in peace agreements.

regional autonomy is already in place. The authors argue that governmental power sharing has the strongest mitigating effect as it requires central and regional elites to coordinate and negotiate policies on a day-to-day basis with each other in lieu of operating through separate compartments of governance. When political influence is shared within the national executive, secessionist activity is better contained (Ibid: 364). Equivalently, Mattes and Savun (2009: 742) argue that integrating regional elites in the governmental system and including them in the state’s central decision-making processes is crucial for them to not develop anti-state incentives.

2.2 Theoretical framework

Building on theories where the role of the state is explicit in the causal argument, the theoretical framework of the thesis differentiates itself from earlier research by suggesting a new orientation. The aim is to attempt to reconcile several inadequacies identified in previous literature on decentralization and secessionism. Firstly, most of the scholarly work done in this field ignore cases where decentralized arrangements have attained positive outcomes by mainly concentrating on cases which have experienced or are currently experiencing violent conflict.

Secondly, research has overlooked on the capabilities of self-rule and shared rule to influence intrastate cleavages and secessionist conflict in different ways (Bakke 2015: 10). Thirdly, seldom has an evaluation of the interaction and combination of territorial power sharing and governmental power sharing been made. Lastly, research strategies rarely allow for the possibility of scrutinizing reverse causation (Cederman et al. 2015: 356).

In contrast to prior research, I propose that in order for decentralization to have a significant preventive effect against secessionism, a high degree of governmental power sharing – or shared rule – is required. Territorial power sharing – or self-rule – might alleviate secessionism, particularly short-term as it offers more political influence than no autonomy, but it is deemed insufficient on its own. Instead, it is argued that a combination of self-rule and shared rule should be the preeminent way to contain secessionist conflict long-term (see Durante 2018).

Most importantly, however, is the capacity of regional governments to influence national decision-making. Regional populations and elites are legitimized by autonomy concessions.

Such arrangements acknowledge their demands which help demonstrate how central

governments are open for cooperation and negotiation with them. Unfortunately, they also

promote tribalism in the sense that ethnic and regional identifications are institutionalized in combination with a strong supply of resources which can be used for mobilization purposes. In essence, decentralization per se induces a double-edged sword effect on secessionism.

The theoretical argument of the thesis is illustrated below in Figure 1:

Figure 1. Theoretical argument

Note: In the figure, x indicates a conditional relationship.

The implementation of self-rule paradoxically encourages secessionist activity. Regional elites

can mobilize public opinion and put pressure on national authority through the influence

obtained from autonomy. As such, I argue that decentralization is only more likely to be

pacifying if high degrees of governmental power is shared in addition. Inclusion for regional

elites within the state’s national decision-making institutions is imperative and works as a

preventive measure, reassuring that opponents of the central state do not engage in prohibited

behavior by interacting with them through peaceful political negotiations within the boundaries

of the law. Strong political ties between central and regional elites decrease the probability of

mobilization and in turn, the degree of public support for secessionism.

This is illustrated in a theoretical framework below in Figure 2:

Figure 2. Theoretical framework

Note: In the figure, (a), (b), and (c) indicate theoretical phases

The model shows how the degree of shared rule is associated with the degree of public support for secessionism. This association goes through a number of phases (i.e. a decentralization process). In an implementation phase, power sharing institutions are established. Through them, regional elites gain enough political influence to pressure the state, effectively introducing a bargaining phase. Here, central and regional actors negotiate on degrees of power sharing.

Ideally, through consensus, a mobilization phase is avoided. The degree of shared rule is the prominent factor that determines whether a decentralization process experiences a mobilization phase or not. This study’s hypothesis outlines the argument:

Hypothesis: Decentralization with a higher degree of governmental power sharing (shared rule)

is associated with a lower degree of public support for secessionism.

3 Method

In the next section, dependent, independent, and control variables are operationalized. Later, the selection of cases is discussed and a controlled comparison between them is illustrated. The methodological approach chosen for this study is then presented, as is the empirical material and data used, and a discussion on their validity and reliability concludes the chapter.

3.1 Operationalization of dependent and independent variables

3.1.1 Dependent variable

The dependent variable is public support for secessionism. The incidence of secessionism within a state is measured by identifying markers that help determine the territorial population’s preferences with respect to secession. Analyzing the results of independence referenda held in each case should therefore be the most logical primary source of data considering the task at hand. However, given the problematic nature of the Catalan independence referenda (i.e. they were declared illegal, had low turnout, and alleged irregularities were reported), support for secessionism is also operationalized by using alternative opinion polls conducted to measure public sentiments regarding independence.

Previous literature has measured and used the strength of regionalist parties to help explain

secessionism (see Brancati 2007; Massetti and Schakel 2017). While being a valid measurement

– particularly in quantitative research settings where a larger sample size is required, it is

considered problematic and subsequently inappropriate for this study. The main reason is that

while regionalist parties explicitly promote independence, they also tend to have other

important points at issue in their political agenda. Thus, a vote for a specific type of regionalist

party is not per se a definite advocacy for secessionist action.

3.1.2 Independent variable

The independent variable is decentralization. For operationalization purposes, it is conceptualized as an event – that is, the establishment of an elected regional tier of government – and as a process – that is, all transfers of powers from the center to regions that might precede and, most commonly, follow the establishment of elected regional governments. This is because decentralization reforms do not exclusively occur once but are also a continuous process (Massetti and Schakel 2017: 433, 447). Therefore, during this process, the degree of territorial power sharing (self-rule) and governmental power sharing (shared rule) can vary over time.

As the ambition is to investigate the effects of decentralization as both an event and as a process, a particularly useful measurement of regional reform over time is the Regional Authority Index (RAI) proposed by Hooghe et al. (2010). This measurement distinguishes between self-rule – authority exercised by a regional government over those who live in the region – and shared rule – authority exercised by a regional government or its representatives in the country as a whole. In RAI, self-rule is operationalized as “the extent to which a regional government has the authority to act autonomously, the scope of its policy competencies, its capacity to tax, and the extent to which it has an independent legislature and executive.”, and shared rule is operationalized as “the capacity of a regional government to shape national decision making.

National decision making is disaggregated across four areas: normal legislation, executive policy, taxation, and constitutional reform” (Ibid: 14).

The Ethnic Power Relations dataset (EPR-ETH) is part of the GROWup project and contributes

with data for an additional measurement of governmental power sharing (shared rule). It does

so by evaluating the access to central state power of ethnic groups active in the world between

1946 to 2017. It uses specific categories of power sharing configurations indicating

theoretically valid and relevant types, including annual data (Girardin et al. 2015).

The first category has four subcategories on group elite positions in central government and is labeled “included into the executive” (Cederman et al. 2010: 99-101; Girardin et al. 2015):

The second category has three subcategories on group elite exclusion from participation in central government and is labeled “excluded from the executive” (Ibid):

3.2 Case selection

The case selection strategy adopted in this study is based on Lijphart’s (1971) variant of a controlled comparison. By relying on logic of experimentation, a researcher can identify causality. John Stuart Mill based his discussion of the five methods of induction on such logic.

The method of choice for this study is called “method of difference”, and it compares cases that differ in outcome. By excluding variables found across all cases, researchers can through the use of logic eliminate them from having causal effects on outcome variance. This is because those variables cannot be attributed to any of the conditions present in all cases (Mill 1843:

455). The most problematic aspect of using such logic is the fact that there are hardly any cases that exist which are equal on all aspects except one. Mill was well aware of this issue, and as such did not recommend its application on complex social phenomena. Lijphart (1971: 688) however, believes that the logic of elimination can be applied successfully in social science research, stating that such an exacting scientific principle is an unnecessary hindrance.

However, one should always strive to achieve this principle as best as possible.

George and Bennett (2005: 240-241) contend that the chosen research design should always be the point of reference when selecting cases. The selection of cases ought to be guided by what is being studied, and this requires clearly defined categories and scope conditions. Furthermore, selected cases should provide adequate control and variation so as to be able to explain circumstances that are of interest to the study in question.

This study’s case selection strategy is illustrated in Table 1.

Table 1. Controlled comparison

2

For detailed coding schemes of each dimension of regional authority in RAI, see Appendix.

Spain and Catalonia were selected because of the difference to the UK and Scotland based on the dependent variable (outcome) and degree of governmental power sharing (shared rule) – whereas all other variables are similar. Spain is formally a unitary state with a decentralized political system established by the Spanish Constitution of 1978 following the transition to democracy. Regularly characterized as “a nation of nations”, Spain consists of ethnic minorities organized territorially as autonomous communities (Colomer 1998: 40- 41). Catalonia, in particular, was selected because it has commonly been described as “a paradigmatic example of a ‘nation without state’ characterized by the importance of ‘dual identities’, with a predominant ‘civic nationalism’ and claims for self-government of non-secessionist character”

(Serrano 2013: 523). Yet, recent developments emphasize a substantial rise of secessionism in Catalan society (Guibernau 2014: 15; Colomer 2017). Most notable manifestations are the Catalan self-determination referenda held in 2014 and 2017, which resulted in an overwhelming majority vote for independence. Both were declared unconstitutional by the Spanish government and have thus not received either national or international recognition. Opinion polls measuring public sentiment on independence suggest results consistent with a majority in favor of secession following the last held referendum.

The United Kingdom and Scotland were selected on the basis of the difference with Spain and Catalonia in regard to the dependent variable (outcome) and degree of governmental power sharing (shared rule) – in combination with similarities on all other variables. The UK is formally a unitary state consisting of four constituent countries with devolved governments.

Scotland, in particular, was chosen because it has limited self-government within the UK, as

well as representation in the UK Parliament. Executive and legislative powers respectively have

been devolved to the Scottish Government and the Scottish Parliament at Holyrood in

Edinburgh since 1999. The political party with the largest number of seats in the Scottish

Parliament is the Scottish National Party (SNP), which actively campaigns for Scottish

independence. SNP enjoyed an overall majority victory in the 2011 Scottish Parliament

election, eventually leading to a proposed Scottish independence referendum, which was

conducted in 2014. Through the Edinburgh Agreement signed by both governments, the UK

government officially recognized and granted the Scottish parliament the necessary powers to

hold an independence referendum. It also received international recognition. The results were

a majority vote against independence from the United Kingdom (BBC News 2014a). SNP’s

overall majority in the Scottish Parliament was subsequently lost in 2016. Following the

referendum, opinion polls measuring public sentiment on independence indicate results consistent with a majority against secessionism, favoring devolution instead.

Both cases present an empirical puzzle. Spain and the United Kingdom are one of the most decentralized and democratic states in the world, yet secessionism is prevalent. Extensive decentralization has been found to decrease the likelihood of ethno-political mobilization and secessionism, whereas weak and medium levels of decentralization increase it (see Miodownik and Cartrite 2010). Furthermore, the correlation between consolidated democracies and absence of internal conflict is well established in contemporary literature (see Hegre 2014). As such, they can be classified as deviant cases. However, while Scotland and Catalonia share similar context, their experience diverge. In the former case, secessionism has not been as dominant as in the latter, as is highlighted by the contrasting results of their respective independence referendum. Furthermore, while the British government recognized the 2014 Scottish referendum as legitimate, the 2014 and 2017 Catalan independence referenda were declared illegal by the Spanish government, resulting in violent conflict and the incarceration of leading Catalan politicians (The Guardian 2019).

Why does the level of public support for secessionism differ between the cases, and why have

the state’s respective central governments reacted so differently to secessionist demands when

there are good grounds to expect that they would be considerably similar? An effective strategy

to help address cases that are not adequately explained by existing theories is to use case studies,

preferably with more than a single case (Lijphart 1971: 692; Levy 2008: 7).

3.3 Control variables

Variables that can have a causal effect on both the dependent and independent variable of interest – public support for secessionism and decentralization respectively, need to be controlled when comparing different cases. When held constant, control variables cannot be the cause for outcome divergence in compared cases (George and Bennett 2005: 428). The variables that are controlled for in this study’s case comparison are: system of government, degree of democracy, ethnic and cultural fractionalization, socioeconomic status, and regional group concentration. Scores for each variable in each case are presented in Table 1.

System of government is included because the type of executive system in a country may affect the likelihood of secessionism. Presidentialism, a system of government in which voters directly elect a chief executive, as opposed to parliamentary systems of government in which the national legislature elects a chief executive, is thought to increase ethnic conflict and secessionism because executives in presidential systems of government are less likely to represent ethnic groups than executives in parliamentary systems of government (Linz and Valenzuela 1994). In parliamentary systems of government, the executive branch may include ethnic groups through coalition governments (Brancati 2006: 671).

Degree of democracy is included because regional groups residing in democratic states should have less resentment towards the central government when their civil liberties are respected, and their political rights are guaranteed (Fearon and Laitin 2003). In strong democracies, regional groups should be better capable of settling disputes with the central government peacefully (Brancati 2006: 671). Therefore, secessionist activity should not be as severe in well- functioning democracies. Research suggests that the risk of experiencing intrastate conflict is lower in stable democratic states in comparison with fragile democracies and autocracies (Hegre 2014: 163). Degree of democracy is measured using Freedom House’s composite score ranging from 1-100, where a higher score reflects a higher degree of democracy.

Ethnic and cultural fractionalization is included because secessionism has a decreased

possibility of occurring in states which are ethnically homogenous, and an increased chance of

developing in states which are ethnically heterogenous. The likelihood increases when most of

the population does not feel ethnically represented by the central state (Lijphart 1977). The

diversification of ethnicity and culture can influence public stability in a negative way by instigating animosity among groups (Horowitz 1985). Diversified scores measuring ethnicity and culture are provided by Fearon (2003) and used in this study to control for ethnic and cultural fractionalization. The definition provided on ethnic fractionalization derives from the odds of two persons belonging to different ethnicity groups by chance selection. Scores are given numerically (between 0 and 1), where largely fragmented states are placed around zero, bipolar states are placed in between, and homogenous states are placed around 1. The definition provided on cultural fractionalization derives from the estimation of how close groups are culturally, by analyzing what languages are spoken. Scores are given numerically (between 0 and 1), where linguistic differentiated states are placed around zero, and linguistic homogenous states are placed around 1 (Ibid: 196, 206-207, 212).

Socioeconomic status is included because economic development is associated with a decreased probability of states experiencing conflict due to the increased ability of governments to handle rebellion (Fearon and Laitin 2003). Economic development also improves education and social welfare, making the population less susceptible to ‘radical beliefs’ (Lipset 1963).

Socioeconomic inequalities are associated with an increased risk of secessionism (Hechter and Levi 1985). The variable is operationalized by using a metric developed by the United Nations called the Human Development Index (HDI). It measures education, life expectancy, and per capita income to assess the levels of social and economic development of states.

Regional group concentration is included because the literature suggests that geographically-

concentrated group are more likely to experience conflict and secessionism (see Toft 2002,

2003; Cornell 2002; Lichbach 1995; Goemans 2006). The variable is measured using the

Minority at Risk (MAR) dataset, which offers scores on group concentration. Data on group

spatial distribution is indicated by scores between 0 and 3 – where 0 is dispersed, 1 is urban or

minority in one region, 2 is majority in one region and others dispersed, and 3 is concentrated

in one region (Minorities at Risk Project 2009).

3.4 Comparative Process Tracing (CPT): a two-step methodological approach

The method of choice for this study is an innovative approach developed by Bengtsson and Ruonavaara (2017) called ‘Comparative Process Tracing’ or CPT. It is chosen because it is well-suited for investigating empirical-historical puzzles concerning two or more cases, as is the case with this study on Catalonia and Scotland. Additionally, as Walt (1997: 15) explains, process tracing “is especially appropriate because the universe of cases is too small for a statistical analysis and the number of independent variables too large for a rigorous application of John Stuart Mill’s ‘method of difference’ ”. Thus, process tracing is chosen as it can ameliorate the limitations of a controlled comparison (George and Bennett 2005: 531-532).

In short, CPT is a two-step methodological approach that combines theory, chronology, and comparison to make general inference possible. Because it has a comparative element, it uses a broad understanding of process tracing that allows both formal testing of hypotheses and narrative approaches. As such, CPT allows for both deductive and inductive reasoning, assumed appropriate for systematic comparisons of complex political processes.

General understandings of cross-case comparison vary. Bennett et al. (2019: 1) argue that some methods are based on the logic of approximating experimental control while other view comparisons as useful for inspiring theory and producing strong tests of theory without automatically providing a distinct logic of inference from process tracing. In this study, process tracing is broadly referred to a research approach that is focused on tracing processes – that is, that looks at how various social and political outcomes are produced by events that result from actors’ actions and interactions, and various contextual factors. For each studied case, the processes leading “from A to B” – in this case from decentralization to a given degree of public support for secessionism – are reconstructed and analyzed through theoretical social mechanisms (see Table 2), which are then compared in conjunction with the theoretical framework (see Figure 2).

Bengtsson and Ruonavaara (2017: 44, 46-47) argue that their approach is fruitful for the

comparative analysis of all types of processes related to political actors, such as

democratization, globalization, conflict and war, and peace building. Such processes can often

be understood in terms of institutionalization and path dependence. As this study focuses on

the process of decentralization, this method is understood to be suitable for the particular

research purpose at hand. Having said that, in contrast with the approach presented by Bengtsson and Ruonavaara, this study uses a slightly less inductive logic, where theory- informed narratives are combined with the formal testing of a hypothesis, and independent and dependent variables are discussed. Bennett et al. (2019: 13) maintain that “finding a middle ground between coherent narrative and highlighting the inferential weight of key pieces of evidence is an important methodological agenda”. In agreement with this sentiment, this study’s methodological modus operandi aspires at such an ambition.

To sum, the two-step methodological approach by Bengtsson and Ruonavaara is based on critical junctures and political focal points, social mechanisms, path dependence, context and periodization, and counterfactual analysis. Said concepts will be further explained, and their methodological relevance to this study’s empirical application will be further discussed below.

3.4.1 Within-case step

The first step in the analysis constitutes identifying and analyzing each case’s historical and empirical characteristics through the identification of critical junctures and political focal points which have contributed to either the path dependence of decentralization or to its change – where the social mechanisms that made decentralization path dependent will be identified, and an analytical periodization of the development of decentralization will be applied to each case.

The concept of critical junctures – an important part of the first, within-case step of CPT – is well-established within literature on path dependence (see Collier and Collier 1991; Mahoney 2000), where it is regarded as decisive in accounting for institutional and political change. This is because periods of continuous development that are sometimes interrupted by ruptures which create new contextual situations are common in historical social science. Collier and Collier (1991: 29) define a critical juncture as “a period of significant change, which typically occurs in distinct ways in distinct countries (or in other units of analysis) and which is hypothesised to produce distinct legacies”. Similarly, Mahoney (2000: 513) discusses critical junctures as transitional situations in which actors are faced with choices that would open a new path.

Deciding on a new path is seen as creating a legacy, a new path dependence, since it becomes

more difficult to return to the initial starting point where alternatives were still available. In this

study, critical junctures which have contributed to either the path dependence of

decentralization or its change will be identified and analyzed.

However, critical junctures are not the only decision-making situations of interest in an analysis with regards to path dependence and CPT. Bengtsson and Ruonavaara (2017: 52) suggest an additional concept which they label political focal point. The reasoning being that other situations where policy and institutions are discussed are as important as critical junctures, as they may instead demonstrate, manifest, and consolidate the path dependence of a direction taken previously. Unlike critical junctures, focal points are always politically salient. A political focal point is thus a decision point where the restricting role of dominant institutions is made explicit. As such, situations where decentralization is explicitly discussed by actors and subsequently institutionalized will be identified and analyzed in each empirical case.

The concept of mechanisms is central in process tracing (see Beach and Pedersen 2013) as well as in the analysis of path dependence (e.g. Mahoney 2000). The concepts’ core definition is debated within said fields, where different definitions are incompatible. Yet, Bengtsson and Ruonavaara (2017: 53) identify general characteristics found in the various ways of using the term. Mechanisms are regular patterns of actions and interactions. They are causally productive, meaning they bring about outcomes. To be useful in comparative research, patterns discovered in one context should also be possible to be discovered in others – i.e. mechanisms should be portable (see Falleti and Lynch 2009: 1145). In a limited sense, they should also be generalizable. CPT allows for an open-ended approach to mechanistic thinking which makes it suitable for comparison between complex processes which must be based on general patterns instead of precise variables (Bengtsson and Ruonavaara 2017: 54). Therefore, CPT is suitable for a comparison made between two cases of separate decentralization processes – where generalization is obtained through abstraction and inference. Empirical observations made in each case are abstracted through identified mechanisms, and later inferred through a comparison between each context.

What social mechanisms of path dependence can be identified theoretically in a decentralization

process? Bengtsson and Ruonavaara (2017: 55) differentiate between efficiency, legitimacy,

and power mechanisms. Each identifies a different mechanism of institutional reproduction –

i.e. how decentralization is self-reinforced, and each suggests a different mechanism of

institutional change – i.e. how decentralization is strengthened (see Table 2).

Table 2. Theoretical social mechanisms in a decentralization process

Note: Author’s adaptation from Bengtsson and Ruonavaara (2017: 55) and Mahoney (2000: 517).

An efficiency mechanism is emphasized by most path dependence theorists and involves actors’

perceptions of economic and other benefits and costs attached to different choices and outcomes. The logic derives from established institutions’ considerable coordination capacities, which are costly to alter. Institutional change takes place when it is no longer in the self-interest of actors to reproduce a given institution (Ibid). A legitimacy mechanism works on the decision- makers’ and the public’s perceptions of existing institutions as legitimate, which contributes to giving these an advantage over alternative institutional arrangements. Institutional change occurs when actors adopt new subjective evaluations and moral codes concerning appropriateness in relation to cognitive frameworks predominant in their societal context (Mahoney 2000: 523-525). A power mechanism works through inclusion and exclusion of actors in decision-making, their varying influence on decisions, and their differing capacity of establishing the decision-making agenda. Political power is distributed by institutions between group of actors, and this distribution makes it beneficial for some actors to keep the status quo, while others challenge it. Institutional change may materialize when reproduction reaches a critical threshold point, after which time self-reinforcement accentuates conflictual aspects between elites and subordinate groups (Ibid: 521-523). This study’s hypothesis highlights the importance of this particular mechanism in its causal argument (see Figure 1).

Process tracing does not necessarily build on theories of path dependence. However, Bengtsson and Ruonavaara (2017: 48-49) argue that CPT is linked to this type of institutional theory.

Therefore, comparing decentralization processes requires an analysis in terms of continuity and

change. The core notion about path dependence can be considered a truism (i.e. history matters),

and some outcomes are such that they cannot be explained by general causes but only by the sequence of events in their specific history. Sewell 2005: (100-101) defines path dependence as a process where “that what happened at an earlier point in time will affect the outcomes of a sequence of events occurring at a later point in time”. The analysis must be strong on historical description. As such, it is crucial to also analyze how history matters in a decentralization process and what mechanisms are keeping its development on said path. What distinguishes Bengtsson and Ruonavaara’s view is that path dependence is seen as a broad and open-ended conception, so as to be more fruitful for comparison. In contrast to a strong and deterministic definition proposed by Mahoney (2000: 511), Bengtsson and Ruonavaara (2017: 50-51) argue for a weak definition, where path dependence is seen as “a historical pattern in which a certain outcome can be traced back to a particular set of events on the basis of empirical observation guided by social theory”. This definition leaves open the possibility of endogenous institutional change and allows for degrees of path dependence, considered important in an analysis of decentralization processes where change is possible through the actions and interactions of actors involved. Whereas a deterministic definition produces a robust explanation, it also demands a ceteris paribus assumption where extreme scope conditions are required for the validity of the generalization – making it less constructive in comparative research on politics.

Nevertheless, Mahoney (2000: 508) claims that processes that are anomalous are useful for analysis in terms of path dependence. As is previously argued, the cases chosen in this study satisfy such ambition, as outcome that are predicted by theory did not occur. Lastly, a theoretical framework is important in a perspective of weak path dependence, where narration is, to a certain extent, theoretically laden. The CPT approach to process tracing is actor-centered and thinly rationalistic. By thinly rational action, Bengtsson and Ruonavaara (2017: 51) mean that actors can be assumed to act on the basis of their beliefs to accomplish a goal. Beliefs and goals do not have to be rational in an “objective” sense – thinly rational actors act for a reason on the basis of these which are largely shaped by their context.

3.4.2 Between-case step

The second step in the analysis involves a systematic comparison of the historical and empirical

characteristics between the cases through the theoretical framework of this study (see Figure

2), and by the theoretical social mechanisms identified in each case.

An integral part in the second, comparative step of CPT is establishing the relevant context as well as deciding and specifying the empirical focus in the reconstruction of the selected cases analyzed in the first step. Accounts of how decentralization with a given degree of shared rule lead to a given degree of public support for secessionism must consider the social, political, and cultural contexts in which actors act and social mechanisms operate. This is because, as Bengtsson and Ruonavaara (2017: 56) allude to, the course of action taken by actors – whether considered rational or “meaningful” – depends on the context in which they find themselves.

As specified in case selection, careful consideration of context and empirical focus was effectuated through a controlled comparison between Catalonia and Scotland (see Table 1).

Evidently, the contexts of both cases are similar enough to make a systematic comparison possible, and different enough to make such a comparison meaningful. This is critical, because in order to make comparison relevant, analogous contexts are required. In open-ended CPT however, identical contexts are not mandatory (Ibid).

Another significant component in the comparative step of CPT is a theoretically informed model of periodization. As Bengtsson and Ruonavaara (2017: 57) argue, a combination of such a model in conjunction with social mechanisms based on thin rationality can serve as a basis for comparing patterns between complex macro-level processes which are characterized by too many operationalized variables and too impregnated with context. Through periodization, temporally ordered processes of decentralization are divided up into subsequent, homogenous eras, periods, or epochs. Within periods, there is similarity, and between periods, there are differences. Distinguishing them is conditioned by the chosen perspective. Like context, periodization should be theory-relevant, to focus the analysis in a way that contributes to comparative consistency. The model created and used in this study is illustrated in Figure 2.

Finally, Mahoney (2000) recommends a type of counterfactual thinking that uses theory to

answer “what if” questions, in order to obtain ideas of what would have happened if the

development had taken a different path than what actually did. Bengtsson and Ruonavaara

(2017: 59) argue that counterfactual analysis can give further evidence to the significance of

path dependence. In this study, the decision was made to disregard such an examination, for the

reason that it is considered less fruitful when only two cases are used, and the limited scope of

this study prohibits a constructive counterfactual analysis to be adequately produced.

3.5 Empirical material and data

The CPT method entails exploring in depth the mechanisms of path dependence in selected cases. As such, access to primary actor material in the form of documents and interviews is required (Bengtsson and Ruonavaara 2018: 55-56). Bennett et al. (2019: 49) argue that

“qualitative research is at its best when it showcases salient evidence in its original form”, for example, through direct quotations from documents and informants. Evidence presented directly permit readers to assess whether the author’s interpretations and inferences are convincing. Therefore, primary sources in the form of official statements, public speeches, and interviews, etc. are prioritized in this study. Critically reviewed secondary sources akin to relevant literature and peer-reviewed scholarly articles are used in combination. Relevant in this case means that research done in the fields concerning the study in question was systematically selected and intentionally included. Furthermore, additional information from non-academic sources including reports, databases and statistics by international organizations and institutions, as well as public data and official documents by governments will be a part of the study’s analysis. Lastly, news media articles will be used as complimentary material.

3.6 Validity and reliability

An analysis on how decentralization impacts the degree of public support for secessionism in established democracies is dependent on using valid and reliable markers that indicate and measure the theoretical concepts relevant to the study at hand. The conceptual definitions previously outlined have been purposely chosen to be suitable for the studied cases, so that variables and mechanisms can be appropriately operationalized. This assists on finding corresponding markers when assessing the empirics. One of the advantages of qualitative case studies is that they allow for more sophisticated conceptualizations, which provide researchers with greater validity degrees when applied on small-n research designs. In comparison, quantitative studies are limited in their conceptual advancement by requiring extensive case samples (George and Bennett 2005: 83-85). In this study, general concepts including

“decentralization”, “power sharing”, and “secessionism” are theoretically refined using descriptions which designate important subcategories through labels such as “territorial”,

“governmental”, “self-rule”, “shared rule”, “inclusion” and “exclusion”, etc.

4 Within-case empirical analysis

In this section, each case is analyzed by an overview of their decentralization and public support for secessionism. An empirical analysis as a narrative for how the process has developed in each phase is then presented, followed by an interpretation. This is repeated in each phase.

4.1 Spain and Catalonia

4.1.1 Overview of decentralization and public support for secessionism

Spain is a decentralized unitary state with a diverse population represented by ethnic minorities territorially organized as autonomous communities. Frequently referred to as the “State of Autonomies”, it features a unique institutional framework of 17 autonomous communities (AACC), each with regional institutions and democratic constitutional statues of autonomy.

There is not yet consensus on whether Spain should be regarded as a de facto federal state or federation (Moreno 1997: 95). For example, Elazar (1991: 228) refers to it as a “federation in all but its name”, whereas Lijphart (1999: 191) classifies it as a “semi-federal” state.

Catalonia’s adoption of its statute of autonomy in 1979 marked the beginning of the region’s self-government through the Generalitat, where it possesses jurisdiction over various areas of competence excluding those which are exclusive to the Spanish government. Catalonia has limited shared rule, where representatives account for a minority of the senate which can influence national tax policy and some reserved powers over constitutional appointments.

However, they can be overridden by a majority in the lower house. Thus, a lack of decision- making in the national arena and no control over the Spanish constitution, is evident (Hooghe et al. 2010: 142, 221-223). RAI scores on self-rule and shared rule are shown in Table 1.

Catalonia and its population have had representation in the national parliament in every

legislature but have never been formally part of the government. Although Catalan politicians

have been ministers in different national governments, they did not represent the interests of

Catalan citizen. Support of Catalan parties of the investiture of national governments did not

mean direct access to executive power, but rather a set of policy concessions or exchange

support at the regional level. As such, the EPR-ETH categorize Catalan regional elites as

excluded from the executive and coded as powerless (Girardin et al. 2015).

Concerning public support for secessionism, the 2014 Catalan self-determination referendum asked two questions: "Do you want Catalonia to become a State?" and "Do you want this State to be independent?”. The second question could only be answered by those who had answered Yes to the first one (Generalitat de Catalunya 2014). Results showed an 80.8% support for the Yes–Yes option, with an overall turnout ranging between 37% to 41.6%, depending on source.

Yes-No received 10.1% and the No option 4.5%. The second Catalan independence referendum held in 2017 asked only one question: "Do you want Catalonia to become an independent state in the form of a republic?", where voters could answer with “Yes” or “No”. 90.09% voted for independence while 7.87% voted against, with a turnout of 42.34% (Generalitat de Catalunya 2017). Following the 2017 referendum, the Center for Opinion Studies (CEO) conducted a total of 5 polls on public sentiments towards independence, where 4 showed a predominant “In favor” answer, and 1 showed a predominant “Against” answer. “In favor” has been the predominant answer on the opinion polls conducted until July 2019.

In addition, CEO has performed regular polls studying opinion of Catalan citizens regarding Catalonia’s political status within Spain. Table 3 illustrates an annual poll using the question:

"Which kind of political entity should Catalonia be with respect to Spain?"

Table 3. Public opinion on Catalonia’s political status within Spain

Note: Author’s elaboration using data from Centre d’Estudis d’Opinió (CEO) 2005-2019.

4.1.2 Implementation phase (a); 1978-1980

The beginnings of secessionism in Catalonia can be traced back to the mid 19th century (Ucelay-Da Cal 2018: 15). The country was led by general Francisco Franco and his authoritarian regime for almost four decades. During this time period, a “sacred unity of the homeland” was enforced. Ideologies with secessionist and federalist tendencies were actively suppressed, as were ethnic identities in the form of cultural and linguistic differences. The Spanish transition to democracy began with Franco’s death in 1975. Following the first free election since 1936, the 1977 Spanish general election was called by Prime Minister Adolfo Suárez to elect a constituent assembly that was to promulgate a new constitution. Formulated in combination with negotiations between central and regional elites – including former Catalan President Josep Tarradellas, who spent 38 years in exile – the fathers of the 1978 Spanish constitution had to strike a balance between the centralist view inherited from Franco, and the pluralistic view of Spain as a “nation of nations” (Moreno 1997). It laid the foundation to the conception known as the “State of Autonomies” – a compromise of "(1) the idea of an indivisible and solely Spanish nation-state and (2) a concept of Spain as an ensemble of diverse peoples, historic nationalities, and regions" (Ibid: 69). The genesis of regional autonomy in Catalonia is illustrated in the Art. 2 of the constitution (Hernández Lafuente 1999: 50-67):

“The Constitution is based on the indissoluble unity of the Spanish Nation, the common and indivisible homeland of all Spaniards; it recognizes and guarantees the right to self-government of the nationalities and regions of which it is composed and the solidarity among them all.”

The new Statute of Autonomy of Catalonia was submitted for a referendum held in October 1979, where the Catalan population voted for a Catalan Parliament with a majority of 88.1% in favor. Following the results, the statute was approved by the Congress in November, and by the Senate in December (De La Granja et al. 2001: 215-216). The first Catalan election was held in March of 1980, which saw the Catalan nationalist electoral alliance Convergence and Union (CiU), under the leadership of Jordi Pujol, receive the majority of votes. As President of the Generalitat, Pujol and his party would lay the foundation of the new regional administration.

He would later reveal his vision of autonomy, declaring that “[it] responds to the need to

institutionally recognize the will of one’s own way of being, with the intention of bringing

power closer to the people. Since the people in Spain are diverse, their autonomies must be

diverse” (Miranda and Castelló 2003).

Later the same year, Suárez would reflect on the institutional changes adopted stating:

“It must be remembered that in three years we have transitioned from an authoritarian regime to a democratic system, several elections have been held, a Constitution has been drafted through compromise, political legislation has been renewed and important reforms have been made to achieve a more fair and stable social life (…)” and “When evaluating the situation, the positives of the Government's actions are very eloquent: Basque and Catalan self-government institutions have been constituted with democratic normality (…)” (El País 1980)

He added that the decentralization reform as a method of gradual change has “advantages and costs” and that the Spanish Government “assume [the costs] without complexes, because we believe the advantages far outweigh them”. Suárez would further argue that:

“Compared to the centralist model, the superiority of strong autonomies is based on the advantages offered by bringing the decision power closer to the governed and identifying the different communities with their cultural and historical being, which does not have to be annulled by the general and common [being] of the entire nation.” (Ibid)

4.1.3 Interpretation

The death of Franco defines a critical juncture in Spanish society, as elite actors inherited a situation of transition following decades of dictatorship. The new path taken during this period of time was manifested through the drafting and promulgation of the 1978 constitution, which established Spain as a democratic state – a framework present to this day. Under this political focal point, the unity of Spain as a nation was declared in accordance with the formal recognition of its plural identity by granting autonomy to its historical nationalities, including Catalonia. Through the approval of a Catalan statute, decentralization was established.

During the implementation phase, key actors recognized that decentralization offered more benefits than disadvantages in comparison to the centralist model, which dominated for nearly four decades. This time period served as a learning process – effectively justifying institutional change, and through cost-benefit assessments, decentralization was subsequently strengthened.

Thus, an efficiency mechanism is evident. It does not account for the developments in its

entirety, however. Equally apparent is the alteration in the values and beliefs of the central

government, which saw decentralization as ‘fairer’ and more democratic. Key actors in general,

and Suárez in particular, recognized decentralized institutions as more legitimate, and more in

line with the new democratic context that comprehensively had transformed Spanish society at